Intracellular Posttranslational Modifications of ... - Journal of Virology

JOURNAL OF VIROLOGY, May 1996, p. 2974–2981 0022-538X/96/$04.0010 Copyright q 1996, American Society for Microbiology

Vol. 70, No. 5

Intracellular Posttranslational Modifications of S1133 Avian Reovirus Proteins ´ N VARELA, JOSE ´ MARTIŃEZ-COSTAS,† MOISE ´ S MALLO,‡ RUBE AND JAVIER BENAVENTE* Departamento de Bioquı´mica y Biologıá Molecular, Facultad de Farmacia, 15706 Santiago de Compostela, Spain Received 22 September 1995/Accepted 13 February 1996

Avian reovirus S1133 specifies at least 10 primary translation products, eight of which are present in the viral particle and two of which are nonstructural proteins. In the work presented here, we studied the covalent modifications undergone by these translation products in the infected cell. The structural polypeptide m2 was shown to be intracellularly modified by both myristoylation and proteolysis. The site-specific cleavage of m2 yielded a large carboxy-terminal fragment and a myristoylated ;5,500-Mr peptide corresponding to the amino terminus. Both m2 and its cleavage products were found to be structural components of the reovirion. Most avian reovirus proteins were found to be glycosylated and to have a blocking group at the amino terminus. In contrast to the mammalian reovirus system, none of the avian reovirus polypeptides was found to incorporate phosphorus during infection. Our results add to current understanding of the similarities and differences between avian and mammalian reoviruses. are aware, there have been no subsequent attempts to resolve these discrepancies, and the glycosylation status of mammalian reovirus proteins remains a matter of controversy. The intracellular posttranslational modifications of avian reovirus polypeptides have been much less completely characterized than those of their mammalian counterparts. In a recent pulse-chase study of infected cells, Ni et al. (28) found a precursor-product relationship between m2 and m2C of avian reovirus 176. The results of a previous study carried out in our laboratory suggest that posttranslational cleavage of the virusencoded primary translation product m2 takes place in avian reovirus S1133-infected chicken embryo fibroblasts (CEF) (46). No other posttranslational modifications of avian reovirus proteins have been described. In the work reported here we carried out an extensive characterization of the intracellular modifications undergone by the primary translation products of the avian reovirus S1133. We found that avian reovirus polypeptides undergo similar posttranslational modifications, except as regards phosphorylation, to those undergone by their mammalian counterparts.

Avian reoviruses are very similar in structure and molecular composition to mammalian reoviruses. Both groups have a genome consisting of 10 segments of double-stranded RNA surrounded by a two-layer capsid which also encloses a number of short oligonucleotides (13). However, avian reoviruses differ from their mammalian counterparts in lacking hemagglutinating activity (11) and in being able to induce fusion of cultured cells (27). Mammalian reovirus type 3, the prototype of the genus Orthoreovirus, encodes 11 primary translation products which can be separated into three size classes: large (l), medium (m), and small (s) (38). The viral polypeptide m1, encoded by the M2 genomic segment, is proteolytically cleaved in the infected cell to yield m1C and m1N, the latter being a small peptide that corresponds to the amino terminus of the precursor. All three m polypeptides are structural components of the virion (30). Intracellular covalent modifications of mammalian reovirus proteins have also been documented. Specifically, m1 and m1N contain an amide-linked myristoyl group attached to their amino-terminal glycines (30), and the presence of this group is necessary for the subsequent site-specific cleavage of m1 to m1C in transfected COS cells (43). There is also evidence for the presence of phosphoserine residues in m1C (16), and both m1 and m1C have been shown to be polyadenylylated and/or ADP-ribosylated (3, 4). The presence of glycoproteins in mammalian reovirions was first suggested by the fact that treatment of purified virions with b-glycosidase, potassium periodate, sodium borohydride, or lysozyme reduced both the hemagglutination titer and the infectivity of the virus (19, 20, 42). However, subsequent work on the glycosylation of mammalian reovirus proteins has produced conflicting results: Krystal et al. reported (15) that m1C was the only polypeptide glycosylated, whereas the results of Lee (18) suggested that all mammalian reovirus polypeptides except s2 are glycoproteins. As far as we

MATERIALS AND METHODS Cells, viruses, and preparation of purified virions. Primary cultures of CEF were prepared from 9- to 10-day-old chicken embryos and grown in medium 199 supplemented with 10% tryptose phosphate broth and 5% calf serum. Strain S1133 of avian reovirus (45) was grown in confluent monolayers of primary CEF. Conditions for growing, purifying, and determining the titer of the virus have been described previously (40). To obtain radiolabeled purified reovirions, 10 mCi of either [35S]methionine-cysteine (Trans 35S-label; ICN; 1,200 Ci/mmol) or [3H]myristic acid (Amersham; 54 Ci/mmol) per ml was added to the cell culture at the onset of the infection. For 32P labeling, 12.5 mCi of 32Pi (ICN) per ml was added 4 h postinfection to cells cultured in phosphate-free medium. All infections were allowed to continue for 30 h at 378C, and in all cases the medium contained 2% dialyzed fetal calf serum. Cells were then harvested for purification of reovirions. Metabolic labeling of infected cells. Monolayer cultures of CEF containing 2 3 106 cells in 35-mm plates were washed with warm Puck’s saline and incubated at 378C for 2 h with 10 PFU of avian reovirus or 0.1 PFU of influenza virus. The unadsorbed virions were then removed (this moment was considered the onset of the infection), and cells were incubated at 378C in a CO2 incubator with medium 199 supplemented with 2% fetal calf serum. At the times indicated for each experiment, cells were incubated at 378C with 25 mCi of [35S]methioninecysteine per ml for 1 h in methionine-free medium or for 6 h in glucose-free medium, with 100 mCi of [3H]glucosamine (Amersham; 30 Ci/mmol) per ml for

* Corresponding author. Phone: 34-81-599157. Fax: 34-81-599157. † Present address: Institute of Virology and Environmental Microbiology, Oxford OX1 3SR, England. ‡ Present address: Department of Developmental Biology, MaxPlanck Institute of Immunobiology, 79108 Freiburg, Germany. 2974

VOL. 70, 1996

PROCESSING OF AVIAN REOVIRUS POLYPEPTIDES

6 h in glucose-free medium, or with 100 mCi of 32Pi for 24 h in phosphate-free medium; the medium in all cases contained 2% dialyzed fetal calf serum. At the end of the labeling period, cultured cells were washed with Puck’s saline and then lysed with 80 ml of cold lysis buffer (10 mM Tris-HCl [pH 8.6], 140 mM NaCl, 1.5 mM MgCl2, 0.5% Nonidet P-40). Cell lysates were scraped off the plate, nuclei were removed by centrifugation, and cytoplasmic extracts were analyzed by electrophoresis, either before or after immunoprecipitation with antibodies to purified virus (1). For pulse-chase analysis, infected cells were incubated with 150 mCi of [35S] methionine-cysteine per ml for 7 min in methionine-free medium and then either lysed immediately or after chasing for the indicated times in medium 199 containing 2% fetal calf serum and 100 mg of nonradiolabeled methionine per ml. In vitro transcription of purified reovirions. Purified S1133 reovirions heat shocked at 608C for 30 s were used to program the synthesis of viral mRNAs as described elsewhere (1), except that nonradiolabeled UTP was used instead of [3H]UTP and that small molecules were removed from the reaction mixture by centrifugation through a 10,000-Mr ultrafree-MC filter unit (Millipore) instead of by exclusion chromatography. Protein synthesis in rabbit reticulocyte lysates. Incubation mixtures (final volume, 25 ml) consisted of 60% lysate (Boehringer Mannheim) plus 1.5 mM magnesium acetate, 100 mM potassium acetate, 25 mM each amino acid except methionine, 1 mCi of [35S]methionine (Amersham; .1,000 Ci/mmol) per ml, 8 mM hemin, 10 mM creatine phosphate, 17 mg of creatine kinase per ml, 0.17 mM dithiothreitol, 0.3 mM CaCl2, 0.66 mM EGTA [ethylene glycol-bis(b-aminoethyl ether)-N,N,N9,N9-tetraacetic acid], and 40 mg of in vitro-transcribed viral mRNA per ml. After 1 h incubation at 308C, samples were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Electrophoretic analysis. Proteins were analyzed by SDS-PAGE on discontinuous gels (17). Several types of resolving gel, with or without gradient, were tested. We selected a 0.75-mm-thick 10% Hydrolink gel (AT Biochem, Malvern, Pa.), which gave good resolution of all viral proteins, including the four m polypeptides. Standard polyacrylamide gels were, however, used in the experiments involving specific staining of 1,2-diol groups and in the investigation of phosphorylation. Low-molecular-weight polypeptides were resolved by electrophoresis as described by Scha¨gger and von Jagow (37). Determination of amino-terminal amino acid sequences. Unfractionated reovirus polypeptides were solubilized by incubating 15 mg of purified reovirions in 0.5 ml of 20 mM sodium phosphate, pH 7.4, containing 2% SDS and 1% 2mercaptoethanol, at 1008C for 3 min. One aliquot (5 mg) was subjected to Edman degradation. Another aliquot (10 mg) was fractionated by electrophoresis on SDS–10% polyacrylamide Laemmli gels, and the separated proteins were transferred to polyvinylidene difluoride paper (Immobilon P; Millipore Corp.) as described elsewhere (29). The electroblot was stained with Ponceau S and destained with water; the bands corresponding to individual proteins were then excised and stored at 2208C. Edman degradation analysis of unfractionated proteins or individual polypeptides was performed in a 477-A gas-phase protein sequencer (Applied Biosystems). Analysis of tryptic digests of m2 and m2C. Protein m2 or m2C (10 mg) was in-gel digested with 1 mg of N-tosyl-L-phenylalanine chloromethyl ketonetreated trypsin for 1 h at 308C, as described by Rosenfeld et al. (36). The resulting peptides were extracted from gel slices as described previously (36) and subsequently separated on a C18 octyldecyl silane Hypersil reverse-phase column (100 by 2.1 mm). The column was eluted with a 5 to 60% linear gradient of acetonitrile containing 0.1% trifluoroacetic acid for 30 min at a flow rate of 0.3 ml/min.

RESULTS Proteolytic cleavage. The results of previous studies in our laboratory suggested that posttranslational cleavage of protein m2, the primary translation product encoded by the S1133 M2 gene, takes place in infected CEF (46). To further characterize this event, we carried out pulse-chase experiments in S1133infected CEF. In Fig. 1, the SDS-PAGE-resolved band pattern of the in vitro translation products of the viral mRNAs is shown alongside that of intracellularly synthesized viral polypeptides and that of purified S1133 reovirions. The 10 viral proteins programmed in vitro by viral mRNAs (Fig. 1, lane 2) were also synthesized in infected CEF during a 5-min pulse with [35S]methionine (Fig. 1, lane 3). However, when similarly pulsed cells were chased for different periods in medium containing nonradioactive methionine, the intensity of the band corresponding to the ;73,000-Mr m2 protein decreased with time, whereas a new band corresponding to an ;68,000-Mr polypeptide emerged and gradually became more intense (Fig. 1, lanes 5 and 7). This new protein was of viral origin, since it was recognized by antibodies to purified virus (Fig. 1, lanes 6 and 8) and since it was present in the reovirion (Fig. 1, lanes 1

2975

FIG. 1. Pulse-chase analysis of [35S]methionine-labeled avian reovirus polypeptides. Protein components of purified S1133 reovirions (lanes 1 and 9) and reticulocyte extracts programmed with in vitro-transcribed viral mRNA (lane 2) were analyzed by SDS-PAGE on a 10% Hydrolink gel. On the same gel were also analyzed cytoplasmic extracts of S1133-infected CEF that had been treated with [35S]methionine for 5 min at 24 h postinfection and then either lysed (lane 3) or incubated with nonradioactive methionine for either 30 min (lane 5) or 90 min (lane 7). Immunoprecipitates of samples in lanes 3, 5, and 7 with antibodies to purified avian reovirus S1133 are shown in lanes 4, 6, and 8, respectively. Band nomenclature follows the conventions adopted for mammalian reoviruses. The Mrs of proteins m2 and m2C were estimated by extrapolation from the known Mrs of mammalian reovirus polypeptides (38) analyzed on the same gel.

and 9) but not in extracts of mock-infected cells (data not shown). These results strongly suggest that this polypeptide is the largest cleavage product of m2, and it was thus designated m2C in accordance with the nomenclatural convention for mammalian reovirus polypeptides (13). Densitometric analysis of the autoradiogram in Fig. 1 indicates that more than 60% of m2 was converted to m2C within the first 30 min of chase and about 90% was converted within the first 90 min (data not shown). Proteolytic cleavage of other primary viral proteins was not detected by these pulse-chase experiments. In order to conclusively demonstrate the precursor-product relationship of m2 and m2C, peptide mapping analysis of both proteins was performed. Gel slices containing m2 or m2C were excised from a stained SDS-polyacrylamide gel and subjected to trypsin digestion (36). The resulting peptides were eluted from the slices and analyzed by reverse-phase high-performance liquid chromatography. As can be seen in Fig. 2A, the trypsin digestion peptides of m2 and m2C display identical chromatogram patterns, confirming that m2C is the major cleavage product of m2. The fact that the minor cleavage product (resulting from the removal of m2C from the precursor) does not show up in the chromatogram for m2 can be attributed to its hydrophobicity (due to the presence of a myristoyl group; see below). We next tried to determine the aminoterminal amino acid sequences of m2 and m2C. After separation by SDS-PAGE, the proteins were transferred to polyvinylidene difluoride membranes and subjected to Edman degradation analysis. No sequence could be obtained for m2, indicating that it has a blocked amino-terminal residue. However, up to 10 amino acid residues could be sequenced by Edman degradation of m2C (Fig. 2B). Given that m2C is the largest cleavage product of m2, our results indicate that the

2976

VARELA ET AL.

FIG. 2. Peptide mapping and sequencing of proteins m2 and m2C. Components of purified avian reovirions were separated by electrophoresis and visualized by Coomassie blue staining. (A) Gel slices containing either m2 or m2C were subjected to trypsin digestion, and the resulting peptides were eluted from the gel slice and subsequently analyzed by reverse-phase high-performance liquid chromatography as described in Materials and Methods. Control, trypsin digestion chromatogram of a gel slice containing no protein. (B) Amino-terminal amino acid sequence obtained by Edman degradation analysis of the SDS-PAGEpurified S1133 polypeptide m2C after electroblotting to a polyvinylidene difluoride membrane.

J. VIROL.

precursor is specifically cleaved near its amino terminus and that m2C is the carboxy-terminal cleavage product of m2. This was confirmed by primer extension analysis of S1133 mRNAs with a degenerate-sequence 20-mer complementary oligonucleotide designed to hybridize to the m2 mRNA fragment that codes for the amino-terminal amino acid sequence of m2C (Fig. 2B). This analysis showed that the codon corresponding to the amino-terminal amino acid of m2C is about 150 bases away from the cap mRNA (data not shown), a distance similar to that expected for the corresponding codon of mammalian m2 mRNA (49), suggesting that mammalian reovirus peptide m1N and avian reovirus peptide m2N have similar sizes. Edman degradation analysis of unfractionated proteins of reovirions gave the same amino-terminal amino acid sequence as in analysis of m2C, suggesting that most structural viral polypeptides contain a blocked amino-terminal residue. Myristoylation. The structural polypeptide m1 of mammalian reoviruses and its cleavage product m1N have been previously found to be myristoylated at their amino-terminal glycine residues (30). To determine whether a similar modification affects S1133 polypeptides, the avian reovirus was grown in CEF in the presence of either [3H]myristic acid or [35S]methionine. Virions were then purified from infected cells, and their labeled components were analyzed by SDS-PAGE and autoradiography (Fig. 3). Two tritium-labeled polypeptides were detected when electrophoresis was stopped before the ion front could exit the gel (Fig. 3A, lane 2). One was a small peptide, and the other was a m-class protein that migrated to the same position as m1 and m2 (compare lanes 1 and 2 in Fig. 3A). To achieve better resolution of the m polypeptides, electrophoresis was allowed to proceed for an additional 2 h after the ion front had reached the bottom of the gel (Fig. 3B).

FIG. 3. SDS-PAGE analysis of proteins from purified avian reovirions. (A and B) Components of avian reovirus S1133 purified from cells incubated with [35S]methionine (lanes 1) or [3H]myristic acid (lanes 2) were analyzed by electrophoresis on 10% Hydrolink gels, and radiolabeled proteins were visualized by fluorography. Electrophoresis was performed at 70 V for 90 min and then at 120 V for 3 (A) or 6 (B) h. (C) Separation of small components of nonradiolabeled virus was performed on a Tricine-SDS-PAGE system as described by Scha¨gger and von Jagow (37). Resolved proteins in the gel were fixed with formaldehyde in ethanol as reported by Steck et al. (41) and subsequently visualized by the glutaraldehyde-silver nitrate staining procedure of Oakley et al. (31).

VOL. 70, 1996


2977

FIG. 4. Intracellular glycosylation of avian reovirus polypeptides. (A) Cytoplasmic extracts of S1133-infected CEF that had been incubated with either [35S]methionine (lane 1) or [3H]glucosamine (lane 3) were analyzed by electrophoresis on a 10% Hydrolink gel and subsequent fluorography (5). The band patterns for the immunoprecipitates obtained by incubation of samples in lanes 1 and 3 with antibodies to purified avian reovirions are shown in lanes 2 and 4. Samples in lanes 1 and 2 were exposed to X-ray film for 12 h, and those in lanes 3 and 4 were exposed for 1 week. (B) Cytoplasmic extracts of mock-infected CEF (lanes 1 and 2), influenza virus-infected CEF (lanes 3 and 4), and avian reovirus-infected CEF (lanes 5 and 6), labeled either with [35S]methionine (lanes 1, 3, and 5) or with [3H]glucosamine (lanes 2, 4, and 6), were analyzed by electrophoresis on a 7.5% Hydrolink gel. The positions of the bands corresponding to actin and to the influenza virus polypeptides P1 to P3 (Ps), hemagglutinin (HA), neuraminidase (NA), and nucleoprotein (NP) are indicated on the left, while the positions of bands corresponding to avian reovirus polypeptides are indicated on the right. Samples in lanes 1 to 5 were exposed to X-ray film for 1 day, and that in lane 6 was exposed for 3 days.

Under these conditions the small peptide was not retained within the gel (Fig. 3B, lane 2) and m1 and m2 were resolved as two distinct bands (Fig. 3B, lane 1), allowing identification of m2 as the tritium-labeled protein (Fig. 3B; compare lanes 1 and 2). To rule out the possibility that interconversion between fatty acids occurred during metabolic labeling of infected cells, we next tried to identify the fatty acid associated with m2 and with the small peptide. To this end, [3H]myristate-labeled reovirions were incubated at 1108C for 12 h in 6 N HCl; the organic fraction was then extracted with toluene and analyzed by reverse-phase C18 thin-layer chromatography, as described elsewhere (30). The 3H-labeled compound obtained after acid hydrolysis comigrated with myristic acid but not with lauric, palmitic, or stearic acid (data not shown), indicating that both m2 and m2N have a covalently linked myristoyl group. The presence of a myristoyl group in both m2 and the small peptide strongly suggests that the latter is the amino-terminal cleavage product of m2, and accordingly, this peptide was designated m2N. No tritium was associated with m2C, a finding which is consistent with it being the product of proteolytic removal of m2N from m2. Densitometric analysis of lane 2 of Fig. 3A showed that avian reovirions contain six times more copies of m2N than of m2 (data not shown), a ratio similar to that found for m2C and m2 in purified reovirions by densitometric analysis of Coomassie-blue-stained gels (data not shown). When purified avian reovirions were electrophoretically analyzed on a gel system specially designed to resolve small peptides (37), m2N was visualized as a ;5,500-Mr band after silver staining of the gel (Fig. 3C). Glycosylation. To investigate the possibility of intracellular glycosylation of the avian reovirus polypeptides, S1133-in-

fected CEF were incubated with [3H]glucosamine and cell lysates were analyzed by SDS-PAGE and fluorography. Figure 4A shows the fluorogram from such an analysis, in which cytoplasmic extracts of [3H]glucosamine-labeled infected cells either before (Fig. 4A, lane 3) or after (Fig. 4A, lane 4) immunoprecipitation with antibodies to purified avian reovirus are compared with extracts from [35S]methionine-labeled cells (Fig. 4A, lanes 1 and 2). Most of the intracellularly synthesized viral proteins, including l1, l3, m1, m2, m2C, mNS, s2, sNS, and s3, were found to incorporate tritium during infection, but strikingly, no tritium was incorporated into protein s1 (compare lanes 1 and 3 in Fig. 4A). The procedure was not sufficiently sensitive to detect the incorporation of tritium into protein l2. All of the tritium-labeled viral polypeptides, with the exception of the poorly immunogenic proteins m1 and mNS, were recognized by antibodies raised to purified reovirions, confirming their viral identity. The absence of tritium in protein s1 suggests that little or no conversion of glucosamine into amino acids took place during the labeling period and, therefore, that the bands in lanes 3 and 4 of Fig. 4A represent glycoproteins radioactively labeled on the sugar moiety. To further clarify this point, uninfected cells or cells infected with either influenza virus or avian reovirus were incubated with [35S]methionine or [3H]glucosamine, under identical conditions, and their cytoplasmic extracts were subsequently analyzed by electrophoresis and fluorography (Fig. 4B). The results clearly show that the proteins labeled in cells incubated with [35S]methionine (Fig. 4B, lanes 1, 3, and 5) are completely different from those labeled in cells incubated with [3H]glucosamine (Fig. 4B, lanes 2, 4, and 6). For example, actin is the most

2978

VARELA ET AL.

FIG. 5. Identification of S1133 reovirion glycoproteins by periodic acid-silver staining. (A) A mixture (3 mg each) of b-galactosidase, phosphorylase B, bovine serum albumin, and ovalbumin was resolved by electrophoresis on two 10% Laemmli gels (17). One gel was stained with Coomassie blue (lane 1), and the other (lane 2) was stained with periodic acid-silver nitrate, as described by Dubray and Bezard (9). (B) S1133 polypeptides isolated from purified virus by phenol extraction and acetone precipitation of the phenolic phase (32) were redissolved by boiling in Laemmli sample buffer prior to SDS-PAGE analysis as for panel A. Gels were stained either with Coomassie blue (lane 1) or with periodic acid-silver nitrate (lane 2).

prominent polypeptide in uninfected cells (Fig. 4B, lane 1) but is not even faintly labeled in uninfected cells incubated with [3H]glucosamine (Fig. 4B, lane 2). A similar situation occurs with other unidentified cellular polypeptides (Fig. 4B; compare lanes 1 and 2). Conversely, several polypeptides that are clearly seen in extracts of cells incubated with [3H]glucosamine are hardly detectable in extracts of cells incubated with [35S]methionine (Fig. 4B, lanes 1 and 2). The situation is still more evident with influenza virus-infected cells. Of all the [35S]methionine-labeled polypeptides (Fig. 4B, lane 3), only hemagglutinin and neuraminidase (24) were labeled when incubated with [3H]glucosamine (Fig. 4B, lane 4). The results obtained with avian reovirus-infected cells again confirm that most avian reovirus polypeptides, but not s1, incorporate tritium on incubation with [3H]glucosamine (Fig. 4B; compare lanes 5 and 6). These results are in agreement with previous findings which indicate that little of the radioactive glucosamine in the cell is converted to non-amino monosaccharides while more than 90% of the radioactivity incorporated from [6-3H]glucosamine ends up in the amino sugars of complex carbohydrates (8, 47). The glycoprotein nature of the viral polypeptides was further investigated by a different approach. This method uses periodic acid oxidation followed by silver staining to detect 1,2-diol groups of protein-bound carbohydrates in polyacrylamide gels (9). For these experiments, proteins were resolved in normal polyacrylamide gels, since some component of Hydrolink interfered with sugar-specific staining. We first tested the validity of the method with a mixture containing two glycosylated proteins, ovalbumin and phosphorylase B (9), and two nonglycosylated proteins, bovine serum albumin (9) and b-galactosidase (7). After separation by SDS-PAGE, four bands corresponding to these proteins could be visualized by Coomassie blue staining (Fig. 5A, lane 1); however, only the bands corresponding to ovalbumin and phosphorylase B were visualized when stained with periodic acid-silver nitrate (Fig. 5B, lane 2), confirming the validity of the method for specific detection of glycoproteins. Periodic acid-silver nitrate staining of the SDS-PAGEresolved proteins of the avian reovirion showed that polypeptides l1, l3, m2C, and s2, but not s1, contained 1,2-diol groups (Fig. 5B, lane 2). Results obtained with overloaded gels also

J. VIROL.

indicated that the minor polypeptide s3 is a glycoprotein (data not shown). Under the electrophoretic conditions used in this experiment, m1 and m2 comigrated as a single band just ahead of m2C. When the electrophoresis was allowed to proceed for a further 2 h, however, protein s3 was not retained within the gel whereas m1 and m2 were resolved as two distinct Coomassie blue-stained bands; the presence of 1,2-diol groups on both polypeptides was confirmed by periodic acid-silver nitrate staining (data not shown). Once again, this method was not sensitive enough to determine whether l2 is glycosylated. These results are in complete agreement with those obtained by metabolic labeling (Fig. 4) and clearly demonstrate that polypeptide s1 is not a glycoprotein. Taken together, these results demonstrate that all S1133specified polypeptides (with the exception of s1 and possibly l2) are glycoproteins and that m2C, s2, and sNS are the most heavily glycosylated. Phosphorylation. The results presented so far indicate that avian and mammalian reovirus proteins are similarly modified in their respective host cells. Since m1 and m1C of mammalian reovirus have been shown to contain covalently linked phosphorus in several different molecular forms (3, 4, 16), we next investigated the possible phosphorylation of avian reovirus proteins. To this end, cytoplasmic extracts of either infected or mock-infected 32Pi-treated CEF were immunoprecipitated with anti-S1133 antibodies and then analyzed by SDS-PAGE and autoradiography (Fig. 6A). For this analysis a 5 to 10% gradient polyacrylamide gel was used, to ensure separation of viral polypeptides from genomic segments. As can be seen from Fig. 6A, 32P was incorporated into all three size classes of genomic segment (lane 2) but not into any of the viral polypeptides (compare lanes 1 and 2). Furthermore, all of the immunoprecipitated 32P-radiolabeled proteins of S1133-infected CEF were also present in their mock-infected counterparts

FIG. 6. Phosphorylation of reovirus polypeptides. (A) Autoradiogram of SDS-PAGE-resolved (5 to 10% gradient gel) components of cytoplasmic extracts of 35S-radiolabeled S1133-infected CEF (lane 1) and of immunoprecipitates of cytoplasmic extracts of either 32P-labeled S1133-infected CEF (lane 2) or 32Plabeled mock-infected CEF (lane 3). (B) Purified reovirions of 32P-labeled mammalian reovirus type 3 (lane 1) and of 35S-labeled (lane 2) or 32P-labeled (lane 3) avian reovirus S1133 were analyzed as for panel A. The positions of avian reovirus genomic segments and proteins (A) and that of mammalian reovirus protein m1C (B) are indicated.

VOL. 70, 1996


2979

FIG. 7. Comparison of the amino-terminal amino acid sequences of the avian reovirus S1133 protein m2C and the mammalian reovirus type 1 protein m1C. The latter sequence is the one reported by Wiener and Joklik (49).

(compare lanes 2 and 3). To confirm that no phosphorylation of avian reovirus structural polypeptides takes place in infected cells, we next purified avian and mammalian reovirions from extracts of 32P-labeled infected CEF and mouse L cells, respectively, and analyzed their radioactive components as for Fig. 6A. Whereas phosphorylation of the m1C protein of mammalian reovirions was evident (Fig. 6B, lane 1), none of the polypeptides of avian reovirions was labeled with 32P (Fig. 6B; compare lanes 2 and 3). These results indicate that incorporation of phosphorus into avian reovirus polypeptides does not occur in S1133-infected CEF. DISCUSSION Evidence presented in this article demonstrates that most avian reovirus polypeptides are posttranslationally modified in infected CEF. Our results also show that most of the posttranslational modifications reportedly undergone by mammalian reovirus proteins (38) are undergone by their avian counterparts. However, avian and mammalian reoviruses appear to differ as regards protein phosphorylation. Thus, whereas mammalian reovirus protein m1C incorporates phosphorus when the virus is grown in L cells (Fig. 6B, lane 1), none of the avian reovirus polypeptides does so when grown in CEF (Fig. 6B, lane 3). This does not seem to be due to a difference in protein phosphatase activity between CEF and L cells, since no phosphorylation of avian reovirus proteins was observed in S1133infected L cells and since m1C remained phosphorylated in mammalian reovirions after incubation with cytoplasmic extracts of CEF (data not shown). The absence of phosphorus covalently linked to the avian reovirus proteins also indicates that they are not bound to oligo(A) or ADP-ribose, both of which have been reported to be linked to mammalian reovirus protein m1C (3, 4). Phosphorylation and desphosphorylation of proteins are important posttranslational events that regulate a variety of cellular processes (6, 14). These modifications also seem to play critical roles in the replication of many viruses. Thus, in several protein kinase-containing viruses, these posttranslational events have been reported to be implicated in the regulation of viral genome assembly and morphogenesis (21, 34) and in the modulation of the activity of the viral transcriptase (21, 23, 25, 35). However, the role that phosphorylation of protein m1C plays in mammalian reovirus replication is largely unknown, and since no endogenous protein kinase activity has been detected in these viruses (39), it has been speculated that phosphorylation of reovirus proteins may in fact have no functional significance (16). Our results showing that there is no phosphorylation of S1133 polypeptides are consistent with this hypothesis. Furthermore, no incorporation of the g-Pi group of ATP into structural viral polypeptides was detected after heat shock incubation or chymotrypsin digestion of purified S1133 reovirions (data not shown); both treatments have been shown to switch on the elongation activity of the viral transcriptase (22). These results and the absence of protein phosphorylation during in vitro transcription of reoviral particles (data not shown)

strongly suggest that the activity of the viral transcriptase is not modulated by phosphorylation. Pulse-chase analysis of infected cells suggested that m2, the primary translation product of the S1133 M2 gene (46), is intracellularly cleaved to produce the slightly smaller polypeptide m2C, and this was confirmed by peptide mapping. This cleavage occurs with less than 100% efficiency, as indicated both by the presence of small amounts of the precursor in purified reovirions and by pulse-chase analysis of infected cells. Furthermore, the presence of a myristoyl group in m2 and in m2N and the fact that m2 has a blocked amino group at its amino-terminal residue strongly suggest that m2N is the aminoterminal fragment of m2 and that cleavage occurs near the amino terminus of the precursor. Several additional observations strongly suggest that m2C and m2N are site-specific cleavage products of m2. First, m2C and m2N are present in approximately the same copy numbers in purified reovirions. Second, the estimated molecular weights of m2 and of m2C plus m2N accord with their being precursor and products, respectively. Third, similar proteolytic cleavages have been reported for the m2 protein of avian reovirus 176 (28) and for the myristoylated m1 protein of mammalian reovirus (30). Fourth, avian reovirus protein m2 and its mammalian counterpart m1 are both specifically cleaved on the amino-terminal side of a proline residue (Fig. 7). Furthermore, these two proteins show a degree of sequence similarity on the carboxy-terminal side of the cleavage site (Fig. 7). Our results agree with the notion that, in most myristoylated proteins, the myristic acid is bound to the free amino group of an amino-terminal glycine residue (44). We are currently investigating whether the myristic acid of m2/m2N is bound to a glycine residue and whether the protein encoded by the M2 gene contains conserved amino-terminal sequences similar to those of other N-myristoylated proteins (i.e., a glycine residue in position 2 and a small uncharged residue in position 6) (30). The fact that myristoylation and cleavage of the M2 primary translation product are conserved features among members of the genus Orthoreovirus suggests that these posttranslational events play important roles in reovirus replication. It further suggests that the M2 gene products are structurally, and possibly functionally, homologous in avian and mammalian reoviruses. The results presented here provide evidence that most S1133 polypeptides are glycoproteins, a situation similar to that previously reported for mammalian reovirus polypeptides (18). Whereas our results clearly demonstrate that the S1133 protein s1 is not a glycoprotein, the glycosylation status of l2 is not clear, since its low immunogenicity and its scarcity in both purified reovirions and infected cells make it difficult to detect this protein by the procedures used in this study. Given the general assumption that most glycosylation takes place within intracellular vesicles, our finding that most S1133 polypeptides are glycoproteins is unexpected, since (unlike with rotavirus glycoproteins VP7 and NS28 [10]) none of the newly synthesized reovirus polypeptides has been reported to enter the endoplasmic reticulum. However, recent studies have

2980

VARELA ET AL.

revealed a novel type of protein glycosylation that takes place in the cytoplasm and yields glycoproteins bearing single Nacetylglucosamine moieties linked via an O-glycosidic bond to serine and/or threonine residues (12). Several virus-encoded proteins have been shown to be posttranslationally modified in this way, including the fiber proteins of adenovirus types 2 and 5 (26), a 149-kDa virion protein of human cytomegalovirus (2), and the v-erbA oncoprotein of avian erythroblastosis virus (33). Experiments to elucidate the nature of the carbohydrates linked to avian reovirus polypeptides are in progress in our laboratory. The absence of glycosylation of mammalian reovirus protein s2 and of avian reovirus protein s1, together with their similar copy numbers and identical intravirion distributions (data not shown), argues in favor of strong structural and functional homology between these two proteins. Finally, two more observations regarding the posttranslational modifications of S1133 polypeptides deserve mention. First, Edman degradation analysis of m2C and of unfractionated reovirion polypeptides gave the same amino acid sequence, suggesting that most structural viral polypeptides are posttranslationally modified with a blocking group at their amino terminus. This situation is similar to that reported for mammalian reovirus proteins (32). Second, extended-run electrophoretic analysis of S1133 reovirions (Fig. 3B, lane 1) and of the in vitro translation products of the S2 gene (data not shown) revealed that protein s1 migrates as two distinct bands. One possible explanation for this finding is that some of the primary translation products of the s2 mRNA undergo a posttranslational modification that affects electrophoretic mobility. Alternatively, it is possible that initiation of translation of the s2 message starts at either of two alternative AUG codons, giving rise to two different proteins, one of which has additional amino acid residues at its amino terminus. We favor the second explanation since in vitro translation of the s2 message produced the same polypeptide doublet (data not shown) and since pulse-chase analysis of infected cells did not provide any evidence of posttranslational modification of protein s1 (Fig. 1). Furthermore, a similar situation has been reported for bluetongue virus protein VP6 (48); the two closely migrating forms of this protein are the result of initiation of protein synthesis at two distinct sites and not of posttranslational modification (48). Future work in our laboratory will address these questions. ACKNOWLEDGMENTS We thank Aaron J. Shatkin and Luis Carrasco for critically reviewing the manuscript. We are grateful to Laboratorios Intervet S.A. (Salamanca, Spain) for providing specific-pathogen-free embryonated eggs. This work was supported by the Programa Sectorial de Promocio ń General del Conocimiento of the Ministerio de Educacio ń y Ciencia of Spain (grant PB91-0780). R.V. and M.M. are recipients of predoctoral fellowships from the Xunta de Galicia. REFERENCES 1. Benavente, J., and A. J. Shatkin. 1988. Avian reovirus mRNAs are nonfunctional in infected mouse cells: translational basis for virus host-range restriction. Proc. Natl. Acad. Sci. USA 85:4257–4261. 2. Benko, D. M., R. S. Haltiwanger, G. W. Hart, and W. Gibson. 1988. Virion basic phosphoprotein from human cytomegalovirus contains O-linked Nacetylglucosamine. Proc. Natl. Acad. Sci. USA 85:2573–2577. 3. Carter, C. A. 1979. Polyadenylylation of proteins in reovirions. Proc. Natl. Acad. Sci. USA 76:3087–3091. 4. Carter, C. A., B. Y. Lin, and M. Metley. 1980. Polyadenylylation of reovirus proteins: analysis of the RNA bound to structural proteins. J. Biol. Chem. 255:6479–6485. 5. Chamberlain, J. P. 1979. Fluorographic detection of radioactivity in polyacrylamide gels with the water soluble fluor, sodium salicylate. Anal. Biochem. 98:132–135. 6. Cohen, P. 1982. The role of protein phosphorylation in neural and hormonal

J. VIROL. control of cellular activity. Nature (London) 296:613–620. 7. Craven, G. R., E. Steers, Jr., and C. B. Anfisen. 1965. Purification, composition and molecular weight of the b-galactosidase of Escherichia coli K-12. J. Biol. Chem. 240:2468–2477. 8. Cummings, R. D. 1993. Structural characterization of N-glycans obtained from metabolically-radiolabelled glycoproteins, p. 243–289. In M. Fukuda and A. Kobata (ed.), Glycobiology: a practical approach. IRL Press, Oxford. 9. Dubray, G., and G. Bezard. 1982. A highly sensitive periodic acid-silver stain for 1,2-diol groups of glycoproteins and polysaccharides in polyacrylamide gels. Anal. Biochem. 119:325–329. 10. Estes, M. K., and J. Cohen. 1989. Rotavirus gene structure and function. Microbiol. Rev. 53:410–449. 11. Glass, S. E., S. A. Naqi, C. F. Hall, and K. M. Kerr. 1973. Isolation and characterization of a virus associated with arthritis of chickens. Avian Dis. 17:415–424. 12. Hart, G. W., R. S. Haltiwanger, G. D. Holt, and W. G. Kelly. 1989. Glycosylation in the nucleus and cytoplasm. Annu. Rev. Biochem. 58:841–874. 13. Joklik, W. K. (ed.). 1983. The Reoviridae. Plenum Publishing Corp., New York. 14. Krebs, E. J. 1994. The growth of research on protein phosphorylation. Trends Biochem. Sci. 19:439. 15. Krystal, G., J. Perrault, and A. F. Graham. 1976. Evidence for a glycoprotein in reovirus. Virology 72:308–321. 16. Krystal, G., P. Winn, S. Millward, and S. Sakuma. 1975. Evidence for phosphoproteins in reovirus. Virology 64:505–512. 17. Laemmli, U. K. 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature (London) 227:680–685. 18. Lee, P. W. K. 1983. Glycosylation of reovirus proteins and the effect of 2-deoxy-D-glucose on viral replication and assembly, p. 193–206. In R. W. Compans and D. H. L. Bishop (ed.), Double stranded RNA viruses. Elsevier Publishing Co., Inc., New York. 19. Lerner, A. M., E. J. Bailey, and J. R. Tillotson. 1966. Enterovirus hemagglutination: inhibition by several enzymes and sugars. J. Immunol. 95:1111– 1115. 20. Lerner, A. M., L. D. Gebb, J. R. Tillostson, M. M. Carruthers, and E. J. Bailey. 1966. Enterovirus hemagglutination: inhibition by aldoses and a possible mechanism. J. Immunol. 96:629–636. 21. Liu, M., and D. T. Brown. 1993. Phosphorylation and dephosphorylation events play critical roles in Sindbis virus maturation. Virology 196:703–711. 22. Martinez-Costas, J., R. Varela, and J. Benavente. 1995. Endogenous enzymatic activities of the avian reovirus S1133: identification of the viral capping enzyme. Virology 206:1017–1026. 23. Masters, P. S., and A. M. Banerjee. 1986. Phosphoprotein NS of vesicular stomatitis virus: phosphorylated states and transcriptional activities of intracellular and virion forms. Virology 154:259–270. 24. McCauley, J. W., and B. H. J. Mahy. 1983. Structure and function of the influenza virus genome. Biochem. J. 211:281–294. 25. Moussatche, N., and S. J. Keller. 1991. Phosphorylation of vaccinia virus core proteins during transcription in vitro. J. Virol. 65:2555–2561. 26. Mullis, K. G., R. S. Haltiwanger, G. W. Hart, R. B. Marchase, and J. A. Engler. 1990. Relative accessibility of N-acetylglucosamine in trimers of the adenovirus types 2 and 5 fiber proteins. J. Virol. 64:5317–5323. 27. Ni, Y., and R. F. Ramig. 1993. Characterization of avian reovirus-induced cell fusion: the role of viral structural proteins. Virology 194:705–714. 28. Ni, Y., R. F. Ramig, and M. C. Kemp. 1993. Identification of proteins encoded by avian reoviruses and evidence for post-translational modification. Virology 193:466–469. 29. Nibert, M. L., and B. N. Fields. 1992. A carboxy-terminal fragment of protein m1/m1C is present in infectious subvirion particles of mammalian reoviruses and is proposed to have a role in penetration. J. Virol. 66:6408–6418. 30. Nibert, M. L., L. A. Schiff, and B. N. Fields. 1991. Mammalian reoviruses contain a myristoylated structural protein. J. Virol. 65:1960–1967. 31. Oakley, B. R., D. R. Kirsh, and N. R. Morris. 1980. A simplified ultrasensitive silver stain for detecting proteins in polyacrylamide gels. Anal. Biochem. 105:361–363. 32. Pett, D. M., T. M. Vanaman, and W. K. Joklik. 1973. Studies on the amino and carboxy terminal amino acid sequences of reovirus capsid polypeptides. Virology 52:174–186. 33. Privalsky, M. L. 1990. A subpopulation of the avian erythroblastosis virus v-erbA protein, a member of the nuclear hormone receptor family, is glycosylated. J. Virol. 64:463–466. 34. Pugh, J., A. Zweidler, and J. Summers. 1989. Characterization of the major duck hepatitis B virus core particle protein. J. Virol. 63:1371–1376. 35. Ransone, L. J., and A. Dasgupta. 1989. Multiple isoelectric forms of poliovirus RNA-dependent RNA polymerase: evidence for phosphorylation. J. Virol. 63:4563–4568. 36. Rosenfeld, J., J. Capdevielle, J. C. Guillemot, and P. Ferrara. 1992. In-gel digestion of proteins for internal sequence analysis after one- or two-dimensional gel electrophoresis. Anal. Biochem. 203:173–179. 37. Scha ¨gger, H., and G. von Jagow. 1987. Tricine-sodium dodecyl sulfatepolyacrylamide gel electrophoresis for the separation of proteins in the range from 1 to 100 KDa. Anal. Biochem. 166:368–379.

VOL. 70, 1996 38. Schiff, L. A., and B. N. Fields. 1990. Reoviruses and their replication, p. 1275–1306. In B. N. Fields, D. M. Knipe, et al. (ed.), Virology. Raven Press, New York. 39. Silberstein, H., and J. T. August. 1973. Phosphorylation of animal virus proteins by a virion protein kinase. J. Virol. 12:511–522. 40. Spandidos, D. A., and A. F. Graham. 1976. Physical and chemical characterization of an avian reovirus. J. Virol. 19:968–976. 41. Steck, G., P. Leuthard, and R. R. Bu ¨rk. 1980. Detection of basic proteins and low molecular weight peptides in polyacrylamide gels by formaldehyde fixation. Anal. Biochem. 107:21–24. 42. Tillotson, J. R., and A. M. Lerner. 1966. Effect of periodate oxidation on hemagglutinating and antibody-producing capacities of certain enteroviruses and reoviruses. Proc. Natl. Acad. Sci. USA 56:1143–1150. 43. Tillotson, L., and A. J. Shatkin. 1992. Reovirus polypeptide s3 and Nterminal myristoylation of polypeptide m1 are required for site-specific cleavage to m1C in transfected cells. J. Virol. 66:2180–2186.


2981

44. Towler, D. A., and J. L. Gordon. 1988. The biology and enzymology of eukaryotic protein acylatin. Annu. Rev. Biochem. 57:69–99. 45. van der Heide, L., J. Geissler, and E. S. Bryant. 1974. Infectious tenosynovitis: serological and histopathological response after experimental infection with a Connecticut isolate. Avian Dis. 18:289–296. 46. Varela, R., and J. Benavente. 1994. Protein coding assignment of the avian reovirus S1133. J. Virol. 68:6775–6777. 47. Varki, A. 1994. Metabolic radiolabeling of glycoconjugates. Methods Enzymol. 230:16–32. 48. Wade-Evans, A. M., P. P. C. Mertens, and G. J. Belsham. 1992. Sequence of genome 9 of bluetongue virus (serotype 1, South Africa) and expression analysis demonstrating that different forms of VP6 are derived from initiation of protein synthesis at two distinct sites. J. Gen. Virol. 73:3023–3026. 49. Wiener, J. R., and W. K. Joklik. 1988. Evolution of reovirus genes: a comparison of serotype 1, 2 and 3 M2 genome segments which encode the major structural capsid protein m1C. Virology 163:603–613.